Integrations
What is predictive maintenance?
Predictive maintenance uses condition data and models to forecast when equipment is likely to fail or drift, so you intervene before the failure — not on a fixed calendar. In buildings that usually means reading BMS points, meters, and sensors you already have. The output should be a ranked list of what to do next, not another dashboard.
How predictive maintenance works
Sense the equipment's live condition, model or forecast when it is likely to fail or drift (with a confidence bound), then prioritise the intervention. Calendar preventive maintenance waits for a date or runtime counter; predictive maintenance waits for the signal that the failure window is opening. The useful artefact is a ranked next-action list, not a wall of charts.
Predictive vs preventive vs condition-based
Predictive maintenance forecasts from condition plus history. Preventive maintenance fires on a calendar or runtime schedule regardless of condition. Condition-based maintenance acts when a live threshold or rule is crossed. The three overlap in practice; the full three-way comparison lives in our condition-based maintenance guide. The table below is the short version.
| Approach | Trigger | Typical use |
|---|---|---|
| Predictive maintenance | Model/forecast on condition + history | Catch drift before failure |
| Preventive maintenance | Calendar or runtime | Stable wear patterns, compliance |
| Condition-based maintenance | Threshold / rule on live condition | Act when a limit is crossed |
What data buildings actually have
Most commercial buildings already produce the feed: BMS points (temperatures, valve positions, runtimes, alarms), energy meters, and optional IoT sensors such as vibration where it exists. New hardware is rarely the first requirement; reading what you already log is. For how that plays out on HVAC and plant in estates, see predictive maintenance for buildings.
Where software fits (and where it does not)
Analytics software that ranks interventions is not a CMMS. A CMMS owns work orders, labour, spares, and PM schedules. Predictive analytics should hand a prioritised queue into that workflow; it should not pretend to be the work-order system. FrostLogic Explore's buildings predictive maintenance capability is that queue layer. For a vendor roundup, see our predictive maintenance software comparison.
FrostLogic Explore
Explore is Sensor Intelligence: it reads the BMS, meter, and sensor points you already run and returns a prioritised maintenance queue with evidence. Explore starts read-only by default; write access is granted one scope at a time after review or auto-run inside that scope. It does not replace a CMMS. See the prioritised maintenance queue for how the product applies this.
Before you ask
Predictive maintenance, the common questions.
Predictive maintenance uses condition data and models to forecast when equipment is likely to fail or drift, so you intervene before the failure rather than on a fixed calendar.
Preventive maintenance runs on a calendar or runtime schedule whether the asset needs it or not. Predictive maintenance times the intervention to a forecasted failure or drift window. For the full three-way with condition-based maintenance, see our condition-based maintenance article.
Usually the BMS points, meters, and sensors you already have: temperatures, runtimes, energy, and vibration where it exists. New hardware is rarely the first step.
No. A CMMS runs work orders, labour, spares, and PM schedules. Predictive maintenance ranks what deserves a work order; it does not replace the work-order system.
No. Explore is Sensor Intelligence: a prioritised decision queue on sensor and BMS data. It does not manage work orders, vendors, or PM calendars.
Explore reads building BMS, meter, and sensor data and returns a prioritised maintenance queue with evidence. Explore starts read-only by default; write access is granted one scope at a time. See the product page for the buildings capability.
Related terms
Keep reading
From glossary to product
Turn building forecasts into a prioritised queue
This page defines predictive maintenance. FrostLogic Explore applies it as Sensor Intelligence: one ranked maintenance queue from the BMS, meters and sensors you already run. Explore starts read-only by default; write scopes are granted one at a time.
See the prioritised maintenance queue